I understand that Dynatrace defines a database by name or schema, vendor, ip and port number. This means when you install an oneagent on that database host (infra mode), dynatrace shows the host and process where the database runs as links. Perfect!
But sometimes this failes, then they are not visible (or only the host is visible).
I understand that the ip number is used on both ends to make the host connection to the database, but for the process I am not sure. I figured that it is done by port number, but I have seen this is not always the case.
Who knows how this is done?
Solved! Go to Solution.
For a Java process the JDBC sensor extracts all the information from the connection URL, which usually contains host (resolved via DNS), port, schema, etc. There might be unknown JDBC URLs or unsupported connection pools, where the agent cannot extract this data and so no link can be determined even when an agent is running on the DB host.